A multi-center case control study will be conducted and is useful for the study of colorectal cancer as it can provide data on risk factors in a relatively cost-efficient manner. This allow an in-depth interview to be conducted examining the dietary pattern in the past, before the disease developed and other possible variables that may affect the relationship between diet and CRC. Exposures of interest include dietary pattern, anthropometry and physical assessment, physical inactivity and biological markers as presented in Figure 1.
Three patient groups will be enrolled, namely patients diagnosed with colorectal adenoma, newly diagnosed CRC patients and a group of healthy participants.
Convenience sampling method
Cases will be selected among patients aged 18 years to 79 years old with neoplastic polyps which include adenomas, traditional and sessile serrated polyps who had visited the outpatient oncology clinic or were admitted into the oncology wards within the period of the study in Hospital Canselor Tuanku Muhriz (HCTM), Hospital Kuala Lumpur (HKL) and Hospital Sultanah Nur Zahirah (HSNZ), Terengganu. It will also involve newly diagnosed patients with primary invasive colorectal cancer at any stage of T0-4, N0, M0 according to American Joint Committee on Cancer (AJCC) TNM system as in Table 1.
| AJCC Stage | Stage grouping | Stage description* |
| 0 | Tis | The cancer is in its earliest stage. This stage is also known as carcinoma in situ or intramucosal carcinoma (Tis). It has not grown beyond the inner layer (mucosa) of the colon or rectum. |
| N0 | ||
| M0 | ||
| I | T1 or T2 | The cancer has grown through the muscularis mucosa into the submucosa (T1), and it may also have grown into the muscularis propria (T2). It has not spread to nearby lymph nodes (N0) or to distant sites (M0). |
| N0 | ||
| M0 | ||
| IIA | T3 | The cancer has grown into the outermost layers of the colon or rectum but has not gone through them (T3). It has not reached nearby organs. It has not spread to nearby lymph nodes (N0) or to distant sites (M0). |
| NO | ||
| M0 | ||
| IIB | T4a | The cancer has grown through the wall of the colon or rectum but has not grown into other nearby tissues or organs (T4a). It has not yet spread to nearby lymph nodes (N0) or to distant sites (M0). |
| N0 | ||
| M0 | ||
| IIC | T4b | The cancer has grown through the wall of the colon or rectum and is attached to or has grown into other nearby tissues or organs (T4b). It has not yet spread to nearby lymph nodes (N0) or to distant sites (M0). |
| N0 | ||
| M0 | ||
| IIIA | T1 or T2 | The cancer has grown through the mucosa into the submucosa (T1), and it may also have grown into the muscularis propria (T2). It has spread to 1 to 3 nearby lymph nodes (N1) or into areas of fat near the lymph nodes but not the nodes themselves (N1c). It has not spread to distant sites (M0). |
| N1/N1c | ||
| M0 | ||
| OR | ||
| T1 | The cancer has grown through the mucosa into the submucosa (T1). It has spread to 4 to 6 nearby lymph nodes (N2a). It has not spread to distant sites (M0). | |
| N2a | ||
| M0 | ||
| T3 or T4a | The cancer has grown into the outermost layers of the colon or rectum (T3) or through the visceral peritoneum (T4a) but has not reached nearby organs. It has spread to 1 to 3 nearby lymph nodes (N1a or N1b) or into areas of fat near the lymph nodes but not the nodes themselves (N1c). It has not spread to distant sites (M0). | |
| N1/N1c | ||
| M0 | ||
| OR | ||
| IIIB | T2 or T3 | The cancer has grown into the muscularis propria (T2) or into the outermost layers of the colon or rectum (T3). It has spread to 4 to 6 nearby lymph nodes (N2a). It has not spread to distant sites (M0). |
| N2a | ||
| M0 | ||
| OR | ||
| T1 or T2 | The cancer has grown through the mucosa into the submucosa (T1), and it might also have grown into the muscularis propria (T2). It has spread to 7 or more nearby lymph nodes (N2b). It has not spread to distant sites (M0). | |
| N2b | ||
| M0 | ||
| T4a | The cancer has grown through the wall of the colon or rectum (including the visceral peritoneum) but has not reached nearby organs (T4a). It has spread to 4 to 6 nearby lymph nodes (N2a).It has not spread to distant sites (M0). | |
| N2a | ||
| M0 | ||
| OR | ||
| IIIC | T3 or T4a | The cancer has grown into the outermost layers of the colon or rectum (T3) or through the visceral peritoneum (T4a) but has not reached nearby organs. It has spread to 7 or more nearby lymph nodes (N2b). It has not spread to distant sites (M0). |
| N2b | ||
| M0 | ||
| OR | ||
| T4b | The cancer has grown through the wall of the colon or rectum and is attached to or has grown into other nearby tissues or organs (T4b). It has spread to at least one nearby lymph node or into areas of fat near the lymph nodes (N1 or N2). It has not spread to distant sites (M0). | |
| N1 or N2 | ||
| M0 | ||
| IVA | Any T | The cancer may or may not have grown through the wall of the colon or rectum (Any T). It might or might not have spread to nearby lymph nodes. (Any N). It has spread to 1 distant organ (such as the liver or lung) or distant set of lymph nodes, but not to distant parts of the peritoneum (the lining of the abdominal cavity) (M1a). |
| Any N | ||
| M1a | ||
| IVB | Any T | The cancer might or might not have grown through the wall of the colon or rectum (Any T). It might or might not have spread to nearby lymph nodes (Any N). It has spread to more than 1 distant organ (such as the liver or lung) or distant set of lymph nodes,but not to distant parts of the peritoneum (the lining of the abdominal cavity) (M1b). |
| Any N | ||
| M1b | ||
| IVC | Any T | The cancer might or might not have grown through the wall of the colon or rectum (Any T). It might or might not have spread to nearby lymph nodes (Any N). It has spread to distant parts of the peritoneum (the lining of the abdominal cavity), and may or may not have spread to distant organs or lymph nodes (M1c). |
| Any N | ||
| M1c |
* The following additional categories are not listed in the table above; • TX, Main tumor cannot be assessed due to lack of information. • T0, No evidence of a primary tumor. • NX, Regional lymph nodes cannot be assessed due to lack of information.
Patient recruitment will be based on the classification made by experts in the field, namely gastroenterologists and surgeons involved in each hospital.
1. Age 18 - 79 years old
2. Not on a special diet regime
3. Able to communicate and carry out the interview
4. Not having any chronic diseases
Case 1
1. Age 18 - 79 years old
2. Patient diagnosed with neoplastic polyps which include adenomas, traditional and sessile serrated polyps
3. Not on a special diet regime
4. Able to communicate and carry out the interview
Case 2
1. Age 18 - 79 years old
2. Newly diagnosed colorectal cancer patient with primary invasive colorectal cancer at any stage except for advanced or metastatic cancer (N1-3, M1)
3. Not on a special diet regime
4. Able to communicate and carry out the interview
Exclusion included those diagnosed with other types of cancer (except colon and rectal).
Allowing an error rate of 2.5%, a level of significance (type 1 error) of 5% and 95% confidence interval and with an a priori estimate of 10% prevalence of CRC in controls and a least odds ratio of 2.35 [26], the OpenEpi sample size calculator indicated that a sample size of 291 (97 participants in each control group, case 1 group and case 2 group) is required to achieve a strong power of 80 % in this study. Taking into consideration the 30% drop outs, we estimate that a sample size of 130 volunteers in each group would be sufficient thus a total of 390 participants for this study.
Self-administered or interview-based questionnaires will be administered to respondents. Questionnaires include socio-demographic variables, information on diet (Diet History Questionnaire with multiple checklists of dietary carcinogen and pro inflammatory diet) and physical activity (International Physical Activity Questionnaire- Short Form). Estimated duration for answering all questionnaires is 30 – 45 minutes.
Anthropometric measurements including body weight, height, waist circumference and body composition will be measured using standard methods and trained researchers. Specific body composition measures will be measured using Inbody Bio analyser.
For novel blood biomarkers, we will collect 10 ml non-fasted blood. Blood plasma will be used for metabolomic analysis and inflammatory markers, while plasma and buffy coat samples will be used for epigenetic marks (global methylation).
A validated version of dietary history questionnaire (DHQ) [27] will be used in this study to determine dietary patterns. Subjects will be interviewed by a dietitian to gather detailed information about their dietary habits including the type, amount, method of cooking and frequency. This includes all foods and beverages taken in their daily routines, including snacks or beverages taken between major meals. In addition, the subjects are also asked about the frequency of having each mealtime, ranging from every day, almost every day (5 to 6 times per week), sometime (3 to 4 times per week), seldom (once to twice per week), very seldom (less than once a week) and never. The time, venue and with whom the meal will be taken and recorded. During diet history assessment, respondents will also be probed using dietary carcinogen / ultra-processed food group and pro inflammatory dietchecklist as in Table 2.
| Pro- inflammatory food | Food that may contain carcinogen/Ultra-processed | |
| Red meat | Sweet / savoury packaged snacks | Noodles / rice vermicelli / kuey teow / laksa |
| Processed meat | Ice cream | Fishballs / meatballs/ vegetable balls |
| Organ meat | Chocolate | Instant noodles |
| Fish, non-fatty | Sweets / Chewing gum | Jelly / pudding |
| Egg | Carbonated drinks | Pizza |
| Sweet drink | Packed bread / bun | Sausage / hotdogs |
| White rice | Margerine / spread | Burger |
| Bread | Biscuits | Nuggets |
| Noodles | Cake and pastry | Chicken / meat stock cubes / instant sauce |
| Margerine | Breakfast cereals | Weight loss / Health supplements / drinks |
| French fries | Energy bar / energy drink | Alcoholic beverages |
| Flavoured milk (Pack) | (Whiskey, gin rum, vodka) | |
| Fruits Yoghurt | ||
| Cocoa drinks | ||
| Fruit cordial / drinks |
Information gathered from DHQ is analysed by using Nutritionist Pro (Axxya Systems Stafford, USA) software. In addition, to detect under-reporting, energy intake was divided by the basal metabolic rate (energy intake / basal metabolic rate).
It is well established that physical activity is one of the factors that convincingly decrease the risk of colorectal cancer as reported by World Cancer Research Fund and American Institute for Cancer Research [18]. In addition, physical activity is also reported to be one of the important factors to improve the prognosis of CRC [28]. In this study, a validated short last 7 days self- administered version of the International Physical Activity Questionnaires (IPAQ) will be used to obtain information on physical activity level.
a. Untargeted Profiling of Serum Samples by High- Resolution Mass Spectrometry
The samples will be processed in random order by mixing 20 µL of serum with 200 µL of acetonitrile; filtered with 0.2 µm Agilent Captiva ND filter plate and analyzed by high-resolution LC–MS as previously described. In short, sample extracts will be injected (2 µL) into a UHPLC–MS system consisting of a 1290 Binary LC system, a Jet Stream ESI source, and a 6550 QTOF mass spectrometer (Agilent Technologies, Santa Clara, CA). A Waters Acquity UPLC HSS T3 reverse phase column (2.1 × 100 mm, 1.8 µm) will be used for chromatographic separation by a water: methanol gradient over a 13-min run. The mass spectrometer will be operated in positive or negative ionization mode and the data were acquired over a mass range of 50 to 1000 Da. Conditions for positive ionization mode will be as follows: drying gas (nitrogen) temperature 175 °C and flow 12 L min–1, sheath gas temperature 350 °C and flow 11 L min–1, nebulizer pressure 45 psi, capillary voltage 3500 V, nozzle voltage 300 V, and fragmentor voltage 175V. Acquisition rate will be set at 1.67 Hz. Conditions for negative ionization mode will be the same except for negative polarity. MS/MS spectra will be acquired using 1.3 Da isolation width with 10, 20, and 40 V collision energies. A quality control (QC) serum pool will be prepared by mixing small aliquots of all study samples; QC samples prepared from the pool will be processed along with the study samples. Blank samples and QC samples will be incorporated into the analytical run at regular intervals to monitor method performance. QC samples will be used to assess overall technical variability through calculation of coefficients of variability for a small subset of known features for all QC samples in the batch. All samples will be injected twice in two separate batches with the mass spectrometer operating in positive and negative ionization modes, respectively.
b. Preprocessing of Mass Spectrometry Data
Molecular features will be extracted from positive and negative ionization mode datasets using Agilent recursive feature extraction workflow as described earlier. In brief, molecular feature extraction algorithms for small molecules will be employed with peak height threshold 800 and 600 counts for positive and negative mode data, respectively, with ions limited to [M+H]+ and [M– H]−. Alignment will be performed with retention time and mass windows of 0.07 min and 15 ppm ± 2 mDa, respectively. Targets for the recursive extraction will be created from features with chromatographic peak heights ≥15 000 and
≥2000 counts in at least 2% of samples for positive and negative mode data, respectively. For recursive analysis, features will be re-extracted with ±10 ppm and ±0.05 min matching tolerance without intensity threshold. Mass will be calculated from ions >10% of peak height and <10% above saturation and final alignment will be performed using the same parameters as above.
c. Identification of Metabolites
Retained features will be first grouped by chromatographic retention time and intensity correlation across subjects, since multiple features often originate from the same metabolite. This will result in several clusters of related masses, which were then annotated cluster by cluster. Mass differences between cluster members will be calculated to search for known fragments and adducts. The m/z values were searched using 8 ppm tolerance in online metabolite databases such as the Human Metabolome Database (HMDB; http:// www.hmdb.ca), the Metlin metabolite database, and the Kyoto Encyclopedia of Genes and Genomes (KEGG).
Annotations will be confirmed by matching of retention time and MS/MS spectra with those of authentic chemical standards wherever possible. The resulting chromatograms were inspected, manually reintegrated where needed, and the resulting peak areas were used as a measurement of intensity for statistical analyses.
Plasma samples obtained from the participants will be aliquoted and stored at 80C for analysis. According to their role as inflammatory biomarkers, circulating factors such as insulin-like growth factor-1 (IGF-1), soluble tumor necrosis factor receptor 2 (sTNFR-2), and interleukin-8 (IL-8) are proposed to mediate the inflammatory response linked to CRC within the tumor microenvironment. Based on the manufacturer’s instructions, all inflammatory markers will be measured using Enzyme-Linked Immunosorbent Assay (ELISA) kits as follows: IGF-1 (Catalog No: ab211651) from Abcam (Cambridge, UK), sTNFR-2 (Catalog No: ELH-TNFR2) from RayBiotech (Norcross, GA, USA), and IL-8 (Catalog No: ELH-IL8) from RayBiotech.
Genomic DNA will be isolated from blood biospecimens using the QIAamp DNA Mini and blood kit (Qiagen Inc., USA) in accordance with the manufacturer’s instructions. DNA quantity and purity will be determined by spectrophotometric absorbance. Global genomic DNA methylation levels will be determined using the 5-mC DNA ELISA kit (Zymo Research, USA) in accordance with the manufacturer’s instructions. The assay relies on a sensitive and specific anti-5-methylcytosine monoclonal antibody which would allow for the accurate quantitation of 5-methylcytosine in DNA samples without the need for bisulfite conversion. Briefly, 100 ng of DNA will be prepared for each assay reaction. DNA will be denatured and applied to the wells of the ELISA plate in the presence of a coating buffer. Samples will be incubated with an anti-5-methylcytosine primary antibody and a horseradish peroxidase conjugate secondary antibody, with the relevant washing steps. Following the addition of a HRP substrate, the quantity of 5-methylcytosine in each sample will be determined based on its absorbance measured at 405 nm using an ELISA plate reader against a series of standards which will be generated using mixtures of the negative and positive control samples included with the kit. Each biospecimen will be assayed in duplicate.
Several machine learning algorithms will be developed, namely neural network (NN), random forest (RF), support vector machine (SVM), neuro-fuzzy (NF), and logistic regression models. The performance of each model will be assessed by cross-validation on test sets of known sample groupings. Following the testing stage, the best method could be applied to new patients, which is expected to return the risk probability of the disease or the exact prognosis of the class (case/control) label. The outcome of the analysis was CRC diagnosis.
The student unpaired t-test and correlation coefficients will be used for mean comparisons. The odds ratios (OR) and 95% confidence intervals (CI) obtained from multivariable logistic regression models will be taken as the measures of predictors of CRC. All statistical analyses will be conducted using IBM SPSS for Windows version 25.0 and statistical significance will be set at p-value of ≤ 0.05.
To date, data collection for this study has been hampered due to COVID19 pandemic. In March 2020, as the COVID-19 pandemic hit Malaysia, universities closed their doors with no information about when they would reopen. A Movement Control Order was implemented by the government. While many undergraduate students
learning, most researchers were asked to carry on with their work from home, and research progress plateaud. Most researchers are affected during this crisis, and this includes data collection in hospitals. Hospital Canselor Tuanku Muhriz UKM, which is the main hospital for data collection, was designated for COVID19 cases and was closed to any research. Patient movements were restricted and data collection for this study was halted for almost one year. This year (2021), cases are on the rise again and another series of Movement Control Order was deployed hence affecting our data collection (Figure 2).
Patients are not keen to be interviewed via phone call and video as our patient demographics encompass those of older age.
When first designing the study, the research team received input from patient representatives who had experienced colorectal cancer in addition to professionals in National Cancer Research Institute. All parties had input into the original grant proposal and played a part in developing and agreeing the research questions, study procedures (eg, inclusion criteria) and choice of primary outcome.
Patients are involved during data collection and at the end of data collection, and they will be called for a group meeting to discuss interpretation of the study. We will also discuss plan of translating the data and data dissemination with the involved stakeholders.